CO2 Pressure Regulator Theory of Operation

[Jim Miller]

I thought it would be good to start a thread on the theory of operation of CO2 pressure regulators, including potential optimizations, potential fault mechanisms, etc.

Rather than start from scratch I found this article to provide a simple explanation of operation to start with:

http://www.docsmachine.com/tech/regs.html

I'll have some thoughts in the next post on optimizations for our use.

jim

 

[Matt F.]

page 1 of this thread has some info regarding the function and differences between dual stage and single stage regs:
http://www.barrreport.com/showthread.php/6470-Dual-Stage-Regulators

 

[Jim Miller]

A balance of forces

On the high pressure (tank) side there is normally 800+psi when the tank is full. This is applied to the closure on the orifice which is likely quite small, a small fraction of a square inch. So the force applied is given by the product of pressure and area.

Opposing this force is the spring connected to the knob we twist to raise the output chamber pressure. When the knob is turned (typically clockwise) the spring is compressed and applies force to the other side of the orifice closure. Once the force applied by the spring is sufficient to overcome the force on the tank side gas escapes from the tank side into the output chamber.

This would continue indefinitely if there was no additional opposing force: the spring exerts a constant force and the tank is exhausting gas and eventually would completely empty.

The reason this doesn't occur is that the output chamber is normally sealed or has an exhaust limiter which causes a pressure buildup in the output chamber. This pressure is applied to a diaphram which has a diameter much greater than that of the orrifice closure.

This pressure on the diaphram, while much lower than the tankside pressure, is applied to a much higher area and is connected to the orifice closure in such a way that it opposes the spring.

When the combination of tank side force and output chamber diaphragm force become greater than the spring force the orifice is again closed.

Note that as the tank becomes depleted of liquid the pressure drops below 800psi and the force on the tank side of the closure becomes less. Since the spring force is constant the only way the closure can be obtained is that the output chamber pressure and resulting force must rise to combine with the lower tank side pressure.

 

[Jim Miller]

Things that can go wrong

This is an inherently mechanical device with friction, stiction and subject to contamination. The closure and the stem connected to the diaphram must move freely as must the spring within its enclosure. If anything rubs in an inconsistent or unexpected way the friction and perhaps stiction will cause inconsistent behavior.

Likewise any contamination the in the orifice will prevent proper closure and increase output chamber pressure.

There are probably others...

 

[Jim Miller]

Potential optimizations

The greater the ratio between the diaphragm size and the orifice size the more it contributes to the offsetting of the spring force. The greater the weight given to output chamber force the more constant the output pressure will be as the tank depletes.

There is a limit to how big you can make the diaphragm before it becomes impractically large. The alternative is to make the orifice smaller. Regulators made for industrial uses such as flooding welding joints can't afford small orifices since they won't pass an acceptable amount of gas without freezing up as the gas expands through it.

In our use we typically want 25gm of CO2 per day assuming a 10lb tank lasts 6 months. This implies that we could use a much smaller orifice than a welder might want. I believe one specification that would describe the ability to pass gass is the SCFM rating or flow coefficient Cv. I think we would want regulators with a large diaphragm and a small Cv rating if such a thing exists.

A dual stage regulator is merely two regulators in series. The first is typically a fixed pressure stage (no knob but perhaps a setscrew) but it operates just as before. The second stage is the one connected to the output chamber we use. The benefit to this is that the second stage sees a rising rather than falling pressure from the first stage as the tank depletes. With a rising intermediate chamber pressure as the input to the second stage, the output chamber actually needs less pressure to offset the spring and will actually fall rather than rise.

Note that none of this is dependent on the quality of the needle valve downstream. The needle valve merely leaks gas from the output chamber at a rate dictated by the size of the opening and the differential pressure across the needle valve.

 

[Jim Miller]

So How Much Do We Use?

The amount of CO2 can vary widely but many report that a 10lb tank provides adequate CO2 when used 8hrs a day and in doing so lasts 6 months.

Breaking this down 10lbs is 4540gms. Divided by 180days gives 25gm/day. The density of CO2 at 0c and 1atm is about 2g/l so that yields 12.5L/day.

Our "day" was 8hr or 480min so the flow rate that our system needs to provide is 26ml/minute.

Breaking this down further lets us compare it to "bubbles per second". This comes down to about 0.5ml/second.

 

[Jim Miller]

Some Victors are pretty good

From the Victor Catalog online some numbers are available on pressure rise for their single stage regulator.

Delivery Pressure Rise
Single stage regulators have an increase in delivery pressure
as the cylinder pressure decreases. Listed below is the amount
of pressure increase per 100 PSIG decrease in cylinder
(inlet) pressure.
SR 4 Series F & G Range 2.4 PSIG
SR 4 Series J & K Range 4.8 PSIG
SR 5 Series 1.0 PSIG
SR 250 Series 0.5 PSIG
SR 350 Series 0.8 PSIG
SR 450 Series 0.6 PSIG
SR 600 Series 0.58 PSIG
SR 700 Series 1.2 PSIG

The change in delivery pressure of a TWO Stage regulator from
full to empty Cylinder (inlet pressure) is negligible.​

Note that the SR250 series only rises 0.5lb for each 100lb decrease in supply pressure. This means if the regulator output is set for 20lbs then by the time the tank drops from 800 to 100psi the output pressure would have only risen 4lbs to 24lbs. Hardly the makings of a dump. Not bad for a single stage regulator!

 

[Left C]

The amount of CO2 can vary widely but many report that a 10lb tank provides adequate CO2 when used 8hrs a day and in doing so lasts 6 months.

Breaking this down 10lbs is 4540gms. Divided by 180days gives 25gm/day. The density of CO2 at 0c and 1atm is about 2g/l so that yields 12.5L/day.

Our "day" was 8hr or 480min so the flow rate that our system needs to provide is 26ml/minute.

Breaking this down further lets us compare it to "bubbles per second". This comes down to about 0.5ml/second.

Is there a specific size of aquarium that you are using as an example? The above calculations do not account for aquarium size. As we know, a large aquarium will require more CO2 than a smaller one.

 

[Jim Miller]

I was picking a typical 75 - 90 gallon tank with a typical RG reactor which most folks find hit their desired number with these sort of CO2 tank duration numbers.

Folks with 180-200g monsters probably aren't worried about plonking down serious money for CO2 equipment.

Jim

 

[Left C]

I remember working on a calculation to determine the amount of CO2 that we use by starting with the formula for the volume of a sphere (V = 4/3¶r^3) to approximate the CO2 bubble volume.

 

[Biollante]

Thoughts?

Hi Jim, All,

Any more thoughts on single stage regulators?

Any thoughts on optimum orifice for our mundane low pressure requirements?

I am one who went a rather more expensive route, though I have wondered why a single stage (less expensive) option should not be possible. :gw

Decades ago I purchased a single stage set-up "designed" for aquariums, though pricy it was poorly engineered and cheaply made.

I have often wondered if the supposed problems with single stage delivery systems had more to do with poor/cheap execution rather than inherent flaw in single stage concept.

Though incoherent and not particularly bright myself; I like reading well thought out pieces as this one.

Biollante

 

[Jim Miller]

I still plan to pursue this line of investigation. Time and money are limiters of course. From their external appearance the made for aquaria designs have either no diaphragms or very small ones. Instead they appear to have a large stem connected to a piston which is acted on by the pressure spring. I think that has a couple of problems.

First, the piston needs a seal for its movement usually an o-ring which by its nature is quite lossy both in friction and static friction, aka "stiction". That will make the response to output pressure flow rather erratic. If you're filling a party balloon or a bicycle tire that probably isn't an issue. The diaphragm needs no such friction-inducing seal since it flexes the small movement amount necessary for opening and closing the poppet/seat closure.

Second, the ratio of areas of the piston and orifice are too close in size to make the piston a large enough contributor in offsetting the spring force. It obviously has an effect since the spring is eventually overcome to maintain the pressure in the output chamber but too small to ensure that it can stabilize the output pressure as the tank pressure falls.

I plan to obtain (as funds permit) a few single stage regulators that appear to be prototypical of the various styles, test them for output stability and publish the results. I also plan to disassemble them to see if I can correlate design features to their performance.

I have two single stage regulators now in hand, one designed "for aquarium use" and another apparently designed as an inflator.

The inflator is a Western Enterprises RP320030. I can't seem to create a direct link to it so http://www.westernenterprises.com/ and wander around a bit until you find it.

I can't seem to find the name of the aquarium one just now but it is German made. I currently have it on a 24oz paintball tank while I finish my 20lb CO2 setup. The output pressure is fixed on this one at 1.5bar which isn't a bad choice but the needle valve must be more like 5* taper and with a too coarse thread pitch for useful adjustment. Since my solenoid is not yet hooked up I use this needle valve as an on/off control with an inline NV-55 as the actual rate controller.

Jim

 

[Jim Miller]

I think the output pressure that many folks use is a contributor to the end-of-tank issues that occur. If the output pressure is set to 10psi then a 20psi increase as the tank empties will result in a total output pressure of 30psi. That would result in a 3x delivery rate of CO2 to their tank. Combine that with a starting water column CO2 concentration of 30ppm at the 10psi setting and the 30psi could potentially be delivering 90ppm gas rates assuming the reactor or diffuser were capable of accomodating such an dissolution rate.

No wonder folks lose fish!

I think a much better approach would be to use a higher output chamber pressure so that a single stage rise would be a much less percentage increase. The SR250 shown above only rises 0.5psi per 100psi input drop. Assuming (always dangerous) this is constant until the tank is empty that represents only a 4lb rise from 800psi to 0psi. That would result in a hardly noticeable delivery rate increase if the output chamber pressure had been set to a nominal 30psi when the tank was full. The SR250 is one of the regulators I'm planning to obtain for testing.

I'm sure that the analog needle valve is the reason why folks use a lower pressure. Most needle valves are built with a 3* taper for ease of manufacture often with a 0.032" orifice and combined with a screw pitch that makes cracking open the needle valve immedidately release CO2 at a rate higher than they would like even at 10psi. At this very low opening the needle valve is touchy to set and often unstable as well.

Using 30psi as an output pressure would make the "cracked open" delivery rate 3x higher.

Using a finer (1*20') taper needle valve and finer (56) pitch screws dramatically increases cost and limits the choice to a single supplier at the current time.

 

[Jim Miller]

Another contributor to end-of-tank issues is the silly practice of trying to squeeze the last bit out of a tank fill. When the liquid CO2 is exhausted, leaving only gaseous CO2, the tank pressure begins dropping from the nominal 800psi indication. If you go through the numbers you find that at this point there is only 10% of the initial CO2 volume left.

Since the refill cost for a CO2 tank 20lbs or less is usually no more that $30 this means that folks are endangering their fish for < $3!!!!

A simple "read the damn gauge weekly" admonition along with "fill as soon as the pressure gauge moves" rule would save a lot of anguish.

 

[Jim Miller]

BTW, this wasn't intended as a soliloquy. I'd like a technical discussion of the inner workings of regulators specifically and the delivery system in general how they might be optimized for our use.

As I've stated before I already have a 2 stage regulator that I built out for my use. The other threads on this forum were a great help in that process and I'm grateful for that. However, the process of sourcing and equipping my regulator convinced me that it is not a viable solution for the vast majority of planted tank enthusiasts.

I don't think that polluting or hijacking that thread for an engineering/optimization discussion would be proper, hence this thread.

I'm not remotely planning to go into business but if I can figure out something that works better I'll be sure to share it publicly so that someone interested could do so.

 

[Jim Miller]

The testing I'm planning will be focused on the end-of-tank output chamber pressure stability. I have several small paintball tanks that I can fill with gas-only from my big 20# tank which will put me right at the interesting delivery point without wasting a lot of CO2 and trips to the refill station. I've purchased a refill accessory for that purpose.

 

[Jim Miller]

A quick ease-of-use optimization for folks who use standard CO2 tanks is obtaining a perma-seal. This is a little brass threaded bit with o-rings on each side that threads onto the CO2 tank and is tightened with a hex key.

The o-rings create a very secure seal for the regulator assembly with much less torque required than a nylon or cardboard crush washer requires. This still isn't as easy as attaching a paintball tank which doesn't require any wrenches but is much easier to obtain a leak-free joint.

Inexpensive nearly everywhere online at only $2.95 and probably last most folks several years.

Highly recommended.

 

[Tom Barr]

This means if the regulator output is set for 20lbs then by the time the tank drops from 800 to 100psi the output pressure would have only risen 4lbs to 24lbs. Hardly the makings of a dump. Not bad for a single stage regulator!

I've never heard of a Victor making an end O tank dump

I've watched this maybe 20-30x as my gas tanks drum out to the bitter end.
Never happened yet.

 

[Tom Barr]

A simple "read the damn gauge weekly" admonition along with "fill as soon as the pressure gauge moves" rule would save a lot of anguish.

Yea, gee, how many times would folks NOT do this?
It's like with EI.........how many folks really test and dose according to the test results?

Not many.........

And only after there's an issue, after the fact......

So this is just good maintenance, but the real world hobbyists is going to over look things.
This is certainly one!

So better to hedge against this.

 

[nipat]

I've never heard of a Victor making an end O tank dump

I've watched this maybe 20-30x as my gas tanks drum out to the bitter end.
Never happened yet.

My Azoo reg has never dumped yet too, after many cylinders (used them until
there was no gas anymore). At least it is reliable, short of those 2 stage things of course.
BTW, I now prefer changing tank when the guage moves approach.